This Airflow environment is used in conjunction with the FOLIO edge-ai to provide Generative AI and other machine learning workflows for FOLIO Library Services Platform.
Run edge-ai locally and create an Airflow HTTP Connection
Follow the directions for running edge-ai locally on your computer and
then create a new HTTP Connection edge-ai with the following parameters:
- host:
host.docker.internal - schema:
http - port: 8000
Create a folio connection with the following parameters:
- host: okapi-url
- login: folio user
- password: folio user password
- json: ```{ "tenant": "tenant_code" }``
To run the embedding DAGs, you'll need to set the following Airflow variable:
- host_root: Full path to the local root directory of
ai-workflows
This DAG processes an incoming JSON instance, typically created by a Generative AI process. It queries the FOLIO system for reference data required by the instance properties. Then, it attempts to match the instance with existing instances in FOLIO by calling edge-ai. If a match is found, the existing instance is returned; otherwise, the new instance is added to FOLIO.